Vous êtes sur la page 1sur 10

VACUUM DISTILLATION METHODS FOR LUBE OILS

INCREASE TURBOMACHINERY RELIABILITY


by
Mark A. Adams
Rotating Equipment Engineer
Amoco Oil Company

Texas City, Texas


and
Heinz P. Bloch
Consulting Engineer
Baytown, Texas

ABSTRACT

Mark A. Adams is a Rotating Equip


ment Engineer in the Central Mainte
nance Division of Amoco Oil Company's
Texas City, Texas, refinery where he is re
sponsible for providing technical support
to the Field Machinist Group.
He received his B.S. degree in Mechani
cal Engineering from Michigan Technolog
ical University, and has served internships
with Caterpillar Tractor Company's Hy
draulics Division and in the Components
Technology Division at Argonne National Laboratory. While at
Argonne he published work on acoustic imaging in breeder reac
tor steam generators.
He has been a Project Engineer and a Maintenance Engineer
for Amoco Oil Company at the Texas City Refinery. Mr. Adams
is a member of the ASME, the Vibration Institute, and Phi
Kappa Phi National Honor Society.

The lube oils in self-contained reservoirs for turbomachinery


in hydrocarbon processing plants are subject to contamination
and deterioration from airborne dust, component debris, sys
tem corrosion, heat, and water. Oil supplies associated with cer
tain gas compression sealing systems are further exposed to po
tential dilution with lighter hydrocarbons, and gaseous dilutants
such as H2S. Excess water and acquired hydrocarbon con
stituents which adversely influence the viscosity and other
characteristics of turbomachinery lube oils must be removed
periodically, if machinery distress is to be avoided.
As part of a determined effort to maintain the reliability of crit
ically important turbomachinery, a major petrochemical com
plex has implemented a lube oil quality assurance program. Vac
uum distillation equipment is used for oil purification, and
periodic analysis of 36 reservoirs is employed to quantifY lube
oil condition. A major refinery in the same geographic area is de
veloping a program \\'ith similar goals.
The relevant experience accrued at both facilities is de
scribed. What is required to have consistently good results and
to obtain the maximum economic benefits is illustrated. This can
be achieved by evaluating and selecting an appropriately de
signed vacuum oil purifier and following up by implementing a
rigorous analysis program.

Heinz P. Bloch is a Licensed Consulting


Engineer specializing in machinery relia
bility improvement.
He holds B.S. and M.S. degrees in
Mechanical Engineering from the New jer
sey Institute of Technology and was
employed by Exxon for more than two dec
ades as a Machinery Specialist. T he pri
mary focus of his technical consulting is in
the areas of maintenance cost reduction
and machinery reliability improvement.
T he depth of Mr. Bloch's involvement in all phases of this work
is indicated in over 40 technical society papers and publications.
In addition, he has authored or co-authored six books and
holds four United States Patents relating to high-speed machin
ery improvements.
Mr. Bloch has been elected to three national honor societies,
is a member of STLE and ASME, and he now serves a broad
range of clients with in depth consultation on reliable and effi
cient operation of high-speed machinery. He is a member of the
International Pump Users Symposium Advisory Committee,
sponsored by the Turbomachinery Laboratory.

INTRODUCTION

Lube oils in self-contained reservoirs at refineries, petro


chemical and gas compression facilities are subject to contamina
tion and deterioration from dust, system corrosion, heat, and
water. In addition, some of the oils in compressor shaft sealing
service may risk being diluted with lighter hydrocarbons and
hydrogen sulfide gas. These contaminants adversely affect the
viscosity, flash point, acidity, and other properties of lube oils
and can lead to higher maintenance requirements, frequent oil
changes, or even machinery failures. Studies have shown that
free or dissolved water in lube oil can significantly reduce the
fatigue life of steel parts [1]. A water content ofO.Ol percent can
reduce the fatigue life of antifriction bearings by as much as 50
percent. Industry experience similarly corroborates the de
leterious effects of water in lube oil on turbomachinery reliabil
ity and component condition [2].
These facts and observations strongly support the contention
that reclamation, oil conditioning, or onstream purification of
turbomachinery lube oils makes economic sense for the over
whelming majority of self-contained lube oil systems found in
41

42

PROCEEDINGS OF THE SEVENTEENTH TURBO MACHINERY SYMPOSIUM

typical hydrocarbon processing facilities, utilities and other pro


cess plants.
PRINCIPAL LUBE OIL
PURIFICATION METHODS EXAMINED

Three principal methods of purifying or reclaiming lube oils,


centrifuging, coalescing and vacuum distillation are typically
found in process plants and utilities in the industrialized coun
tries. Centrifuging and coalescing can remove almost all traces
of free water, but cannot separate emulsified water, dissolved
water, light hydrocarbons or harmful gases such as H2S. The lat
ter is capable of leaching copper out of Babbitt bearings and cer
tain Babbitted oil film seals. Copper backed shoes of tilting pad
bearings are especially susceptible to H2S corrosion. A coating
of copper sulfide forms as the copper corrodes, limiting the heat
transfer properties of the bearings.
Centrifuging consists of separating the high specific gravity
component (water) from the low specific gravity component
(oil). The configuration of the centrifuge bowl must be such that
the overflow port is at the exact level, for a given specific gravity
of lube oil, at which there is no carryover of water into the re
claimed oil or loss of centrifuged oil into the free water being
removed. Experience shows that the complete removal of free
water is rarely achieved by even the best of centrifuges, which
sometimes create a tight oil/water emulsion. While it has been
claimed that with close control of specific gravity almost all free
water could be removed, such close control appears nearly im
possible in compressor installations which use the lubricant as a
seal oil and thus experience absorption of gas [3].
Coalescers are successfully used at many airport installations
to dehydrate jet fuel and other aviation fuels, because they are
subject to extremely close quality control and fuel properties de
viate very little, if at all, from specifications. Free water is the
only contaminant, which makes it possible to have the precise
coalescer cartridge configuration for efficient water removal.
The fibers to which the water clings can be woven and the con
figuration designed for optimum water separation. Both compo
nents to be separated, water and fuel, remain constant in
properties.
Because the oil rarely exhibits constant properties, an investi
gation of purification methods for lube oils in hydrocarbon pro
cessing plants will quickly lead to vacuum distillation, also called
vacuum purification. This purification method is the most suita
ble for a wide range of process plant applications since it is capa
ble of removing free, emulsified and dissolved water, entrained
air, H2S and other gases, and light hydrocarbon fractions up to
C12 chains. Specific gravity has no effect on separation using this
method. At the chemical plant under study, two identical vac
uum oil purifier (or oil conditioner) units have been used regu
larly in this service since the plant was commissioned in the late
1970s [4].
VACUUM PURIFICATION PRINCIPLE AND
EQUIPMENT DESIGN OP TIONS

Vacuum oil purifiers work on the principle of simultaneous ap


plication of vacuum and heat to cause vaporization of contami
nants. A typical vacuum oil purifier is shown in Figures 1 and 2
and schematic diagrams are given in Figures 3 through 9. The
approach taken by the first plant, which will be called the chem
ical plant is represented in Figures 1, 2, and 3.
In Figure 3, which also illustrates the operating principle of
the equipment shown in Figures 1 and 2, the contaminated lube
oil entering the vacuum vessel is controlled by a solenoid valve
which allows a falling thin film and a standing reservoir in the
vacuum vessel. In this particular model, the fluid is exposed to

Figure 1. Vacuum Oil Purifier, Front View (Courtesy Allen Filters).

Figure 2. Vacuum Oil Purifier, Rear View (Courtesy Allen Filters).


,------b.fP R V

INLET

+-

OUTLET

......

* FLOW REDUCING VALVE CONTROLLED BY FREON PRESSURE PRECLUDES


COMPRESSOR OVERLOAD WHEN CONTAMINANTS EXCEED DESIGN PARAMETERS.

Figure 3. Schematic Diagram of Vacuum Oil Purifier Employing


Baffled Aluminum Trays.

70-80C (160-180F) temperature and 1.0 to 25 mm Hg pressure


(29.88 to 28 in Hg vacuum) in a tank-like vessel, while flowing
over baffied, inclined, aluminum trays. As the fluid flows over
the trays, dehydration, deaeration, and degasification are ac-

43

VACUUM DISTILLATION METHODS FOR LUBE OILS INCREASE TURBOMACHINERY RELIABILITY

complished. The vapors are drawn from the vacuum vessel


through a refrigerated condenser (evaporator) by a high vacuum
pump. The condensed vapors settle in a distillate collection
tank, which is drained intermittently, by automatic means. The
non-condensibles are discharged through the vacuum pump
exhaust, either to atmosphere or to a low pressure flare line.
Experience to date with the vacuum oil purifier depicted ln
Figures 1 through 3 has shown it to be efficient and generally
low in maintenance. Overall, it incorporates a minimum
number of parts and features nonexpendable baffled trays
within the vacuum vessel. Reliability and energy efficiency are
achieved in this model with refrigerated condensing, electric oil
heaters which do not contact the oil, and, if economics warrant,
heat exchange between the cooler incoming oil and the warmer
discharging oil. This single-stage vacuum design is typically
equipped with a programmable logic controller and annunciator
system.
At the second plant, which we will call the refinery, the vac
uum oil purifiers used are represented by Figure 4. These units
typically consist of an immersion-type electric heater, vacuum
vessel and dispersion elements for surface area extension. The
oil level inside the vacuum vessel is maintained by a float oper
ated valve which allows the vacuum system to draw oil in as re
quired. An outlet oil pump returns reclaimed oil from the vac
uum vessel to the reservoir. Vacuum is maintained by an eductor
system which uses water from a closed loop system as the motive
fluid. Equipment required for the vacuum system are a water
pump, reservoir, eductor and piping. Light hydrocarbon or
water vapors which are removed from the vacuum vessel are
entrained with the water passing through the eductor and are
condensed or vented off from the circulating water reservoir. An
ambient temperature water condenser is located between the
vacuum vessel and the eductor to condense a portion of the hy
drocarbon vapors and water, thereby reducing the vapor loading
on the eductor. These units were chosen primarily for their low
initial cost, minimal number of components, and apparent
simplicity.

A variety of designs (Figures 5, 6, 7, 8, and 9) illustrate proposals


for a specific user's application [5]. A Roots (lobe-type) vacuum
pump (shown in Figure 5) upstream of the condensing chiller,
could have a negative impact on initial cost and maintenance.
The design also requires an additional vacuum pump and driver.
The dispersion (coalescer) cartridges within the vacuum vessel
may require frequent replacement and sometimes limit the

CONDENSATE
TANK

INLET

.......

OUTLET

......

....______
...
__........

__

OUTLET
PUMP

Figure 5. Vacuum Oil Purifier Incorporating Blower Pump Up


stream of Chiller.

,..----tPRV

INLET

.......

AMBIENT
CONDENSER

INLET

+- --r-+r---'

OUTLET
PUMP

HEATER

PRV

Figure 6. Vacuum Oil Purifier Without Condenser.

PRV'JEa-----1

CONDENSATE
TANK

OU TLET
._
........
.._
.
.... __....__
OUTLET
PUMP

Figure 4. Vacuum Oil Purifier Used at the Study Refinery.

With an increasing number of users opting for vacuum oil


purifiers to reclaim lube oils, a number of manufacturers are en
tering the marketplace with competing executions and config
urations. Each of these may offer certain advantages and/or dis
advantages. An engineering appraisal of several configurations
is presented here. Potential users similarly must evaluate the ad
vantages and disadvantages of all systems under consideration.

VACUUM PUMP
(ROTARY PISTON)

SOLENOID
VALVE

OUTLET
PUMP

Figure 7. Design With Ambient Condenser Downstream of


Blower Pump.

44

PROCEEDINGS OF THE SEVENTEENTH TURBOMACHINERY SYMPOSIUM

AMBIENT
CONDENSER

detrimental at his operating conditions. A high pressure drop


could allow the oil to accumulate in the vapor space in the top
of the vessel, and subsequently carry over into the vacuum line.
The screw type discharge pump in this execution would be sub
ject to cavitation under high vacuum conditions even with low
to moderate water contamination.

INLET
-+-

Figure 8. Two-Stage Vacuum Oil Purifier Diagram.


range of oils which may be processed. European and North
American users have reported instances of additive depletion
when using dispersion cartridges in vacuum service on compres
sor sour seal oils. Some of these users noted an emulsion gradu
ally forming and collecting on the outer surface of the cartridge
through which the oil had passed.
It is suspected that there might be a direct relationship be
tween emulsion formation and additive depletion. No additive
depletion has been experienced by the chemical plant in nine
years of using baflled tray-type vacuum oil purifiers, however,
at vacuum levels of 1.0-10 mm Hg and temperatures of 80 to
85C. This is of interest because a concern is raised occasionally
that high vacuum oil purifiers may remove additives by vaporiz
ing them with the contaminants.
Immersion heaters located upstream of the filter may have
shortened life due to contaminants or lube oil constituents car
bonizing on the surface of the elements. Additionally, high sur
face temperatures (or hot spots) on the elements tend to oxidize
the oil on contact.
The execution shown in Figure 6 lacks either a chiller or am
bient condenser-the subject of discussion later. The reader
may recognize it to be a lower-cost version of the earlier Figure
5, and the same comments would apply. It is also noteworthy
that the rotary vane vacuum pump used here may require fre
quent vane replacement, due to wear and degradation by some
types of contaminants, and deserves the attention of personnel
involved in selecting this equipment.
The design depicted in Figure 7 places the condensing com
ponents downstream of the primary vacuum pump. This makes
the vacuum pump unnecessarily large and subjects the pump in
ternals to potentially harmful degradation from water droplets
and other carryover. Note the ambient condenser to be dis
cussed later, and a condensate purge design which requires an
additional pump and driver.
Examining Figure 8, find a two-stage configuration requiring
a total of seven pumps and drivers which could prove both ex
pensive and complex. Thermodynamics considerations with a
single closed-loop cooling system used here both as an ambient
condenser and as a sealing fluid for the liquid ring vacuum pump
would require close investigation. Again, the user's review effort
would have to focus on these issues, since reliability and main
tainability weigh heavily in a modern plant environment.
Finally, the reviewer would have to consider that Figure 9
might have the potential drawbacks described in the earlier dis
cussions of rotary vane vacuum pumps and immersion heaters,
although here the heaters are preceded by a filter. The crinkled
wire mesh used in this execution creates a pressure drop across
the vacuum vessel which the reviewer must ensure will not be

Figure 9. Design Using Crinkled Wire Mesh.

REMOVING GASES BY
VACUUM DISTILLATION METHODS
Sour seal oil contaminated with C1 to C12 hydrocarbons, hy
drogen, H2S, and HCl can be very effectively degassed by vac
uum methods that are an inherent feature of vacuum purifica
tion units. As illustrated earlier, oil entering the system is
brought to optimum temperature and reduced in viscosity to
promote thin-film flow across the fibrous media contained in the
dispersion cartridge or, for reduced maintenance, over slanted
baflles or trays. As the thin film of oil flows through the vessel,
water, light hydrocarbons and other volatile contaminants are
thus distilled, removed by the vacuum equipment and recon
densed for disposal. The filtered, degassed and dehydrated oil
is then returned to the lubrication or seal oil system via the dis
charge pump.
Just how weli vacuum purification units will degas the oil de
pends on the vessel design, temperature, condensing efficiency
and the degree of vacuum maintained in the vessel. Test and
analysis data from the refinery, the chemical plant and from
capable vendors show that units which effectively raise the tem
perature of the oil, create a sufficient vacuum, condense vapors
at relatively cold temperatures and extend the surface area with
out promoting oil carryover \1\.'ill satisfactorily reclaim severely
contaminated seal oil. In one test, a 19 liter (five gallon) sample
of ISO Grade 43 lube oil containing 75 wppm of H2S and two
percent light hydrocarbons was processed, in a single pass,
through a vacuum purifier unit operating at 26.92 in Hg (76. 2
mm Hg). The reclaimed oil was analyzed and its residual H2S
concentration conservatively estimated at 0. 03 wppm (the vapor
in the sample bottle contained seven mppm of H2S). The
amount of light hydrocarbons was reduced to 0. 25 percent in the
same purification step. A second sample processed through a
vacuum purifier unit operating at approximately 29.8 in Hg (3
mm Hg) was degassed even more effectively, with the vapor con
taining a residual H2S concentration of two mppm. Again, the
amount of light hydrocarbons was reduced from two percent to
0.25 percent in one vacuum purification stage.
The refinery also reports interesting results. Here, the exist
ing units are designed to operate with a vacuum of28 in Hg (48.8

45

VACUUI\1 DISTILLATION METHODS FOR LUBE OILS INCREASE TURBOMACHINERY RELIABILITY

--+--

210

45.6

38.8

180

32.0

150

120

25.3

>

90

18.2

60

10.5

>

20

30
--

Figure 10. Viscosity Improvement During Processing Train A


Oil.

mm Hg) and an inlet oil temperature of 71 oc (l60F). In the


field, vacua above 27 in Hg .(74.2 mm Hg) are commonly
achieved, though the vacuum in purifiers which process sour
seal oil that has not been adequately degassed in degassing
drums has been observed to deteriorate to 21 in Hg (226.6 mm
Hg) or less. Inlet oil temperatures typically range from 60C to
82C (140F to 180F). The vapor pressure/temperature relation
ship of certain lube oil contaminants is such that high degrees
of vacuum at reasonable temperatures are required for effective
purification [6]. It did not, therefore, come as a surprise that
with the lube oil contaminants in the C6 to C10 range, the more
elementary vacuum purifiers could not give consitently good
results. Therefore, batches of unprocessed sour seal oil were
subjected to special testing.
A drum of sour seal oil was collected fiom each of three com
pressors for use in the test: a crude unit wet gas machine (Train
A), an isomerization unit hydrogen machine (Train B), and a re
fi:mner recycle machine (Train C). Sour seal oil contaminants by
weight percent were identified by gas chromatography in the c4
to C10 range in the Train A and Train B samples, and C:3 to C10
in the Train C sample as shown in Table 1. Additionally, the Train
A sample contained H2S gas, nitrogen and carbon dioxide.
!\-larked improvements were noted in viscosity and total acid
number of the test oils as they were processed by multi-pass cir
culation through the vacuum oil purifier. As shown in Figure 10,
the viscosity of the Train A oil improved from 4.5 to 41.0 centi
stokes at 38C (40. 7 to 188.2 SUS at l00F) after several passes.
It was not possible to determine the exact number of passes,
however, because the sample volume was drastically reduced
from approximately 200 liters to 70 liters (55 gallons to 18 galTable 1. Composition of Sour Oil from T hree Process Compres
sors by Weight Percent.
Component
Type

457-J sour
Oil Composition

J-1 Sour
Oil Composition

412-J Sour
Oil Composition

CJ

o.o

o.o

0,13

c4

l. 78

0.28

0.67

c5

22.20

3.39

2.55

c6

26.49

5.60

10.46

c7

10.73

4.90

3.78

c8

4.61

0,13

9.76

C9

0.24

0.01

l. 48

c1o

0.01

0,01

0.67

3 3.9 4

85.68

70.50

c11+

Ions) as the light hydrocarbons were removed. Similarly, the vis


cosity of the Train B oil improved from .31.8 to 56.8 centistokes
at 38C (149.2 to 263.2 SUS at l00F); and the Train Coil, from
4.1 to 28.9 centistokes at 38C (.39.3 to 136.6 SUS at l00F). No
comparison is made here to original oil specifications, since the
testing was limited by time, and because the oil taken for testing
from the Train A and Train B compressors was a mixture of two
viscosity grade oils.
All the test oil samples showed a reduction in total acid
number following purification. The most significant decline in
total acid number occurred in the processing of the Train B oil:
from 0.513 to 0.170 KOH mg/gm of oil.
The effectiveness with which contaminants are removed by
vacuum purification is illustrated in Table 2, which is a compari
son of C3 C10 contaminant content of the Train C test oil before
and after purification. Table 3 was developed from a gas
chromatograph of the distillate collected during the test. The
C3, C4 and most of the C5 contaminants, as gases, were not in
the distillate sample having been exhausted through the vac
uum pump. However, as noted earlier, the C6 to C10 contami
nants were effectively flashed off and recaptured for subsequent
disposal due to the combination of high vacua, adequate temper
ature and refiigerated condensing during the test.
Table 2. Composition of Train C Sour Oil by Weight percent Be
fore and After Purification.
Contaminant
Before

Final

Change

c3

0.13

0.0

0.13

C4

0.67

0.0

0.67

c5

2.55

o.o

c6

10.46

.rype

0.14

2.55
10.32

c7

3. 78

0.16

3.62

c8

9,76

2.73

7.03

Cg

1. 48

0.83

0,65

c1o

0.67

0.41

0.26

Table 3. Composition of Train C Distillate by Weight Percent.


Contaminant Type

Weight %

cs

0.02

c6

5. 23

c7

11.91

ca

7 2. 23

Cg

9. 53

c1o

1. 68

ell+

0.02

Test data from the refinery samples correlate positively with


similar tests conducted by other refineries. Viscosity improve
ment indicated in Figure ll is on an ISO VG 32 compressor seal
oil from a wet gas compressor in another refinery, which had not
been mixed with other oils. It shows viscosity improvement
from 22.6 to 30.0 centistokes at 40C (109 to 141.3 SUS at 100F)
after just 6 passes through a vacuum oil purifier at process condi
tions of 65C to 78C (150oF to 173oF) and l.7 to 15 mm Hg.
Flash point improved during the same test from 38C to 204.5C
(l00F to 400F).

46

PROCEEDINGS OF THE SEVENTEENTH TURBO MACHINERY SYMPOSIU M

::E

150

170

[)?.

0..
0..

32.0

,.

ill
:;1

ffi

...J

NEW OIL
140

0
!!;

29.6

130

27.5!:!!

cl.. 120

25.3

>

!<w
u

,..

;:;

23.0

iii 110
0

>

v.:

'

'

1-'
z

"'

1z

0
u
"'
"'
1"

70
eo

k%

...J

"
!!;

f::::
f/

00

40

1:'::
f::":

lZ

:;:
>

20.6

100

0
30

10
211.5

20
29.1

30
28.7

AO
28.3

50
27.11

60
:!7.8

70
27.2

MM I IN
4
NUMBER OF PASSES

Figure 11. Viscosity Improvement During Processing ISO VG 32


Sour Seal Oil.
CONSIDERATIONS FOR
REFRJGERATED CONDENSING

Vacuum oil purifiers should, in most cases, incorporate refrig


erated condensing means between vacuum vessel and vacuum
pump. The refrigerated condensing system improves both the
processing and energy efficiency of the vacuum purifier. Refrig
erated condensing prevents large quantities of water from going
through the vacuum pump. It is absolutely necessary to have re
frigeration condensing if there are any condensible hydrocar
bons in the system. The condensible hydrocarbons also dissolve
in the vacuum pump lube oil and lower its viscosity. This causes
rapid deterioration of lube oil which can lead to distress and fail
ure of the vacuum pump.
As regards the degree of water removal achievable with vac
uum oil purifiers operating at different levels of vacuum and
temperature, Figure 12 will be of interest. With the final ppm
of water in the lube oil calculated on the basis of Henry's law, it
is clear from the figure that if the vacuum level is above 50 mm
Hg, it is no longer possible to reduce water content in lube oil
to within the desired 40 to 50 ppm range. These findings were
corroborated by field experience reported by vacuum oil
purifier users. There is no significant difference in overall de
hydration efficiency if the system is operated at 10 to 25 mm Hg
vs 2.0 to 10 mm Hg, but the higher vacua may be required for
removal of light hydrocarbons above C8 chains.
REQUIRED P ROCESSING TIME FOR WATER
CONTAMINATED OIL CAN BE CALCULATED

In a typical turbomachinery lube oil reservoir, the bulk oil is


contaminated with a combination of free, emulsified and dis
solved water. Analytical studies and field experience show that
even under the best circumstances, lube oil drain headers and
reservoirs are saturated with moist air [1, 7]. The systems are
usually vented to atmosphere. Temperature differences and cy
clic variations in delta T between vent areas and ambient tem
peratures promote condensation. The possibility of ingesting
wet or contaminated air exists also at the shaft seals. Large
amounts of oil draining from the bearing area back to the reser
voir are known to create suction effects or slightly lower pres
sure regions in the bearing housing. This promotes the inflow of
ambient air through labyrinth seals and, together with conden
sation in reservoirs and vents, explains the fact that even motor
driven turbomachinery experiences lube oil contamination [3].
While nitrogen blanketing is occasionally used in efforts to
exclude atmospheric air, experience shows small quantities of

80
28.8

90
26.4

100
28

110
25.8

120
25.2

130
2-4.8

140
24.4

1!50
24

HG

Figure 12. Effect of Vacuum Level on Performance at Different


Temperatures.

nitrogen to be rather ineffective. It has often been noted that


nitrogen blanketing can promote a false sense of security that
water intrusion is eliminated, especially in high humidity loca
tions. To be effective, large volumes of nitrogen would have to
be used and the procedure would not be economically viable or
justifiable.
A conservative rule of thumb developed at the Chemical
Plant showed that for typical water concentrations in the vicinity
of 1,000 to 10,000 wppm it would be possible to reduce the level
of water contamination down to 40 wppm of dissolved water
after perhaps five "turnovers." In this context, "turnover" would
be calculated by dividing the contents of the lube oil reservoir
(gallons) by the processing rate of the lube oil conditioner (gal
lons per hour). Thus, one would normally keep a 300 gph vac
uum dehydrator hooked up to a 3,000 gallon reservoir for 50
hours:
3000 gallons
300 gallonslhr
T
Where
T
W

and

Q
iw
:l,

W
=

X 5
ln

50.

w
:

processing time, hours


contents of lube oil reservoir, gallons
purifier capacity, gallons per hour
initial water content, decimal expression
final water content, decimal expression

Or, in our example


T

3000
300

1000

40

10 In 25

32 Hours

This would assume a processing efficiency of 100 percent,


whereas a more realistic, generally achievable efficiency would
be 80 percent. Note, also, that the above equation does not take
into account the small amount of water which theoretically con
tinues to enter the reservoir from whatever source.
LUBE OIL ANALYSIS P ROGRAMS

Research efforts spearheaded years ago by the utilities indus


try have led to optimized lube oil analysis methods for steam tur
bine lube oils. An authoritative book on this subject recom
mends testing for color, foreign solids, neutralization number,
_viscosity and water content [8]. However, more recent studies

VACUUM DISTILLATION METHODS FOR LUBE OILS INCREASE TURBO MACHINERY RELIABILITY

indicate that these tests alone are not sufficient for early deter
mination, if oxidation has progressed to an undesirable degree
[9]. Lube oil oxidation can result from prolonged exposure to at
mospheric oxygen, high bearing and reservoir temperatures, or
possibly even excessive heating during processing in vacuum oil
purifiers with incorrect temperature settings.
The chemical plant opted for an analysis program which
checks for appearance, water, flash point, viscosity, total acid
number, and additive content. These tests are further described
in the literature [10, 11].
PURIFICATION AND ANALYSIS:
LONG TERM RESULTS

During the first full year of the onstream lube oil purification
and analysis program, the chemical plant was primarily in
terested in establishing the "before vs after" results of analyses
for dissolved water and oxidation inhibitor. The water check
would demonstrate acceptable overall operation of the two vac
uum oil purifiers, checking the oxidation inhibitor content
would tell us if the lube oil had aged, or if, perhaps, the applica
tion of heat and vacuum in the unit had caused the phenolic oxi
dation inhibitor constituent to be removed from the lube oil.
Four turbine driven boiler fan lube oil systems were found to
be typical of the water contamination analysis results. Each of
these 2006 liter (530 gallon) reservoirs was purified during 12hour shifts and showed "before vs after" water contents of 317/
135, 138/11, 324/47, and 351162 wppm, respectively. The reser
voir with the 135 wppm residual water content was given another
12-hour purification treatment and all four subsequently ear
marked for 18-hour conditioner hookups in the future.
Analyses for depletion of the phenolic oxidation inhibitor
were made in similar "before vs after" fashion on all of the 36
lube oil reservoirs at the chemical plant and also by simultane
ously withdrawing an oil sample from both inlet and outlet con
nections on the vacuum oil purifier. There were no significant
differences in the levels of inhibitor concentrations in the sam
ple streams. Also, no significant differences were found in the
inlet vs. outlet samples when the vacuum oil purifier processing
temperature was experimentally raised to 93C (approximately
200F). In fact, the oxidation inhibitor concentration level has
never dropped below the minimum acceptable level of 0.2 per-

47

cent in any of the 36 lube oil reservoirs during almost nine years
of operation! It can certainly be concluded that vacuum oil
purification at the processing conditions given above does not
result in the removal of desirable additives, at least not from tur
bine lube oils marketed by major lube oil refineries. No effort
was made to define which academic values of temperature and
vacuum would cause concern, but users who theorize that their
oil additives may be highly volatile could perform a series of
mass transfer calculations to establish "safe" vacuum oil purifier
operating conditions.
At the chemical plant, oil leakage from compressor shaft seals
is returned directly to the reservoirs. The authors are aware
however, of many user locations such as refineries and gas com
pression facilities onshore and offshore which collect large quan
tities of seal oil directly from the traps. These oils are then batch
processed through vacuum oil purifiers at the same operating
levels of vacuum and temperature as at the chemical plant with
excellent results. One such skid mounted unit incorporating
three 2,500 liter (660 gallon) batch process tanks and three vac
uum oil purifiers is shown in Figure 13.
The chemical plant examined the restoration of flash point by
comparing "before vs after" analyses with, values expressed in
degrees Fahrenheit: 405/412, 412/426, 397/419, and 415/423. As
expected, vacuum oil purification had resulted in the removal
of light hydrocarbons. Very similar results can be reported for
seal oil systems with oil charges that had occasionally been ex
posed to contact with H2S containing gas streams. In three years
of observation by one of the writers, the oils continued to show
flash point values around 210C (410F). In all of these cases, the
accompanying change in lubricant viscosity was marginal and
appeared to be within the anticipated error band for the kinema
tic viscosimeter which was used in this lube oil analysis.
Finally, the chemical plant was observing very minor upward
changes in total acid number for oils after conditioning. How
ever, all of the 36 reservoirs remained well within the specified
allowable maximum TAN of 0.3 KOH mglgm of oil during the
entire nine year period from 1978 until 1988. In fact, the chem
ical plant is still using the original 83,500 liters (22,000 gallons)
of lube oil which were loaded into the reservoirs when the facil
ity was commissioned in 1978. The plant intends to have the
same oil in place during the next decade.

Figure 13. Batch Process Type Sour Seal Oil Reclaimer Package (Courtesy Allen Filters).

48

PROCEEDINGS OF THE SEVENTEENTH TURBO MACHINERY SYMPOSIUM

BENEFIT-TO- COST RATIO CALCULATED

The benefits of on-stream reclamation or purification of lube


oil have been described in numerous technical papers and other
publications. Monetary gains are almost intuitively evident and
additional documentation contained in Jacobson's and Sullivan's
articles [12, 13], and others can be consulted by engineers in
terested in reliable plant operation.
Two hours of unscheduled downtime brought on by contami
nated lube oil in a major turbocompressor would cost many
plants more than the estimated purchase price which can typi
cally range from $30,000 to $80,000 for small, fixed base to large,
mobile vacuum oil purifiers. Although this fact alone should be
sufficient justification for most programs, a detailed investiga
tion was conducted of annual costs and savings for the inhouse
analysis portion of plantwide turbomachinery lube oil recondi
tioning and analysis program performed at the chemical plant.
These costs consisted of laboratory technician wages and ex
pendables such as glassware and chemicals. Savings included
avoidance of higher charges for outside contract laboratory
work, and labor and electric power savings from the elimination
of "precautionary reclamation" practiced in the many instances
where delayed reporting or logistics problems deprived us of
timely feedback. The net annual savings substantially exceeded
the cost of acquiring supplemental laboratory instrumentation
specifically required for inhouse analysis of lube oil, and not
othernise needed by the laboratory. The chemical plant achieved
a benefit-to-cost ratio approaching 1.8, with discounted cash
flow returns exceeding 100 percent.
Another way of calculating the approximate total savings at
the Chemical Plant from 1979 until 1988 would be as follows:
Costs

Collection and processing of samples, $8,450/yr.


Nine-Year Total:
Incremental analytical instruments:
Vacuum oil purifier initial cost:
Avoided cost for individual (dedicated) purifiers:
Vacuum dehydrator maintenance:
Vacuum dehydrator operating labor (160 hrslhr):
Nine-Year Total:

$ 76,000
6,000
60,000
(68,000)
20,000
36,000
$130,000

Avoided Costs-Method I
(frequent changes prevent machinery downtime)
Four complete oil changes, 4 X 22,000@ $2.49/gal $219,000
Two non-turnaround, scheduled, major turbotrain shutdowns for the purpose of changing
300,000
lube/seal oil
Labor and disposal costs, 36 reservoirs,
4 exhange events @ $900 each
130,000
Nine-Year Avoided Cost Total: $649,000
Avoided Cost-Method II
(infrequent changes cause some machinery downtime)
Two complete oil changes, performed during
plant turnaround 2 X 22,000 gal@ $2.49/gal
$110,000
Labor and disposal costs, 36 reservoirs,
2 exchange events@$900 each
65,000
One turbotrain outage and repair event in 9 years
500,000
(One event per 45 train-years)
Nine-Year Avoided Cost Total: $675,000

In considering the above calculations, prospective users


should substitute their own figures appropriate to the installa
tion. For example, handling and transportation costs could add
400 to 1,000 percent to the cost of oil on an offshore production

platform in the North Sea vs a refinery location in the U.S.


where the oil is manufactured.
CONCLUSION

Both the long term field experience of a modern chemical


plant and recent sample studies performed by a major refinery
have shown that vacuum oil purifiers designed to provide the
essential oil temperature increase, surface area extension and
high vacuum exposure can remove process contaminants, free
water and dissolved water. These optimally designed purifiers
have the ability to restore lube oils to their original properties.
On the subject of cost justification, one knowledgeable source
has calculated that for the majority of lubrication systems using
more than 200 liters (approximately 50 U.S. gallons) oflubricant,
oil analysis generally proves more profitable than a routine time/
dump program [12]. Similar findings have been reported by
large-scale users of hydraulic oil whose reconditioning efforts
have proved successful and profitable [13). Field experience by
a major chemical plant and cost justification studies by both the
chemical plant and a major refinery fully support this contention.
There are several reasons why thoughtful engineers should
make an effort to put in place lube oil preservation and waste
reduction programs; of these the economic and environmental
reasons are most important. A conscientiously implemented
program of lube oil analysis and reconditioning can rapidly pay
for itself through lube oil savings and reductions in machinery
failure frequency.
Lube oil analysis techniques are relatively easy to understand
and automated laboratory equipment makes the job more pre
cise and efficient than a few decades ago. Employing these tech
niques in conjunction with a well designed vacuum distillation
type lube oil conditioner allows prospective users to justify lube
oil reclamation as a considerably better alternative than selling,
burning, or otherwise disposing of lube oil in a modern plant
environment.
REFERENCES

1. Cantley, R. E., "The Effect of Water in Lubricating Oil on


Bearing Fatigue Life," ASLE Trans., 20, (3), pp. 244-248
(1977).
2. Wilson, A. C. M., "CorroSion of Tin Base Babbitt Bearings
in Marine Steam Turbines," Trans. oflnst. of Marine Eng.,
73, (ll) (Discussion) (1961).
3. Bloch, H. P., "Criteria for Water Removal From Mechani
cal Drive Steam Turbine Lube Oils," Lubr. Eng., 36, (12),
pp. 699-707 (1980).
4. Bloch, H. P., "Vacuum Oil Conditioner Removes Contami
nants From Lubricating Oil," Chemical Processing, pp. 8485 (March 1982).
5. Bloch, H. P., and Amin, A. "Optimized Vacuum Purifica
tion Methods for Lubricating Oil," Proceedings of 6th Inter
national Colloquium Industrial Lubricants-Properties, Ap
plication, Disposal; Technische Akademie Esslingen, West
Germany Ganuary 1988).
6. Steele, F. M., "Seal Oil Degassing In Gas Turbine Cen
trifugal Compressors," ASME Paper 84-GT-214, Presented
at International Gas Turbine Conference, Amsterdam
(1984).
7. Bloch, H. P., "Reclaim Compressor Seal Oil," Hydrocarbon
Processing, pp. ll5-ll8 (October 1974).
8. "Oelbuch," Anweisung fuer Pruefung, Ueberwachung und
Pflege der im Elektrischen Betrieb verwendeten Oele
und synthetischen Fluessigkeiten mit Isolier- und

VACUUr..I DISTILLATION METHODS FOR LUBE OILS INCREASE TURBO MACHINERY RELIABILITY

49

Schmiereigenschaften/Herausgeber: V DE\V. 4. Auflage,


Frankfurt: Verlags- und Wirtschafi:sgesellschaft der Elek
trizitaetswerke (V\VEW) (196.3).

11. Bloch, H. P., and Geitner, F. K., Machinery Failure


Analysis and Troubleshooting, Houston, Texas: Gulf Pub
lishing Company, pp. 196-205 (1983).

9. Grupp, H., "Moderne Untersuchungsverfahren Fuer


Schmieroele und Hydraulikfluessigkeiten von Dampfturbi
nen," Der Maschinenschaden, .55 Heft 2, pp. 84-87 (1982).

12. Jacobson, D. \V., "How To Handle Waste Oil," Plant Eng.,


p p. 107-109 (June 1982).

10. Bloch, H. P., "Results of a Plant-Wide Turbine Lube Oil Re


conditioning and Analysis Program," Lubr. Eng. 40, (7), pp.
402-408 (1984).

13. Sullivan, J. R., "In-Plant Oil Reclamation-A Case Study,"


Lubr. Eng., 38 pp. 409-411 (July 1982).

50

PROCEEDINGS OF THE SEVETEENTH TURBOMACHINERY SYMPOSIUl\1

Vous aimerez peut-être aussi